24 January 2024 – Day 3 Agenda

Event Agenda: 24 January 2024 – Day 3

Time

Agenda

08:30 - 09:00

Registration

09:05 - 09:10

Day 2 Wrap-up

09:10 - 09:50

Bringing useful quantum computing to the world

More than seven years back we have made quantum computing available through the cloud for the first time. Since then, a continuous stream of technical breakthroughs and improvements on the hardware as well as on the software and algorithm level has been demonstrated to build up a full stack quantum computer. Currently, they are conquering a place in the data and high-performance computing centers. We have stepped up from IBM Quantum System One to IBM Quantum System Two enabled by significant technical achievements, have pushed scale, quality, and speed to the next level, developing a quantum-centric supercomputing architecture and performant software capabilities. For example, we increased the scale to an enormous number of 1121 qubits on a single quantum processor and showed a path to further scaling by modularity. Concomitantly, speed and quality have been increased to boost overall performance. In this presentation I will give an overview of what we have achieved across the quantum computing stack and share our extended roadmap to further advance the Era of Quantum Utility.

09:50 - 10:30

Quantum Optimization: Potential and Challenges

Quantum computing could revolutionize numerous areas in business and science, with optimization often named as a prime candidate to profit from such a revolution. This follows from the fact that optimization problems are ubiquitous, and therefore, any improvement over the state-of-the-art classical algorithms with quantum computers could have a huge impact. Moreover, these improvements could occur across multiple dimensions, such as solution quality, solution diversity, time to solution, and cost to solution. But just how tangible are the benefits of quantum optimization? Grover’s Search algorithm and Quantum Adiabatic Annealing provided a promising start, but it has become clear that more tools are needed to demonstrate near-term benefits in optimization problems. We give an overview on the state of affairs, especially looking at NP-hard combinatorial optimization problems, like, e.g., portfolio optimization.

10:30 - 11:00

Coffee Break

11:00 - 11:30

Quantum-enhanced machine learning using processors engineered with atomic precision

Physical reservoir computing is a machine learning technique that uses the fixed dynamics of some complex dynamical physical system as a mapping to a higher-dimensional feature space. Reservoir computing is expected to enhance the financial sector by capturing temporal dependencies in historical data, facilitating accurate predictions, risk assessments and decision-making that consider the sequential nature of market dynamics. Due to their robustness to noise and relative ease of implementation, reservoir computing implemented in quantum systems are an exciting prospect for near-term intermediate-scale quantum processors, with theoretical studies outperforming equivalently sized classical reservoirs stemming from the expanded size of the computational Hilbert space. In this talk we will show how to leverage atomically precise manufacturing of atoms in silicon to realise quantum enhanced machine learning. By working in two control regimes, we can implement both an extreme learning machine and a reservoir computer with fading memory. These results indicate an exciting avenue for achieving a potential quantum advantage using near-term analogue systems for financial machine learning tasks.

11:30 - 12:00

Quantum cryptography: keeping secrets over the long term

Cryptography is an essential but invisible component of our modern world, keeping private information private and safeguarding our communications in an increasingly digitized landscape. Accelerating development in the field of quantum computing threatens to compromise many of the encryption mechanisms in use today. In this talk I will outline the ways in which we can use quantum technologies to secure our networks against unlimited computational power, and showcase the systems and devices which my team at TII is building to achieve this.

12:00 - 12:30

Quantum computing and quantum communications in the financial industry

The talk describes the state of the art of quantum computing for finance, focusing on the research work conducted by the quantum computing team at JPMorgan Chase in the area of quantum algorithms and applications for financial use cases. We will highlight recent theoretical and experimental advances in quantum optimization and machine learning, including the recent demonstration of a quantum scaling advantage for the Quantum Approximate Optimization Algorithm (QAOA).

12:30 - 13:30

Lunch

13:30 - 14:00

Towards a quantum geometric framework to understand generalization in quantum machine learning

Generalization is the ability of machine learning models to make accurate predictions on new data by learning from training data. Recent work uncovered that uniform generalization bounds, the common approach to understand generalization, often fail to explain quantum machine learning. Here, we introduce the data quantum Fisher information metric (DQFIM) to understand generalization. For variational learning of unitaries, the DQFIM accurately quantifies the circuit parameters and training data needed to successfully train and generalize. Surprisingly, a constant number of training states can already be sufficient for generalization, showing that quantum machine learning can succeed even for very few data. Finally, we uncover two counter-intuitive phenomena in quantum machine learning: First, while symmetries have been believed to improve learning, we show that data without symmetries can be better for generalization. Second, we find that out-of-distribution generalization, where one trains with data drawn from a different distribution than the test data, can be better than using the same distribution for both training and testing. Our work provides a powerful framework to explore the power of quantum machine learning models.

14:00 - 14:30

Preparing for the Most Complex Upgrade Ever

This talk will discuss how to prepare for the most complex upgrade ever, as the Quantum threat is forcing us to upgrade all cryptography systems globally. We can’t leave anything out, and it takes all of us working together to do so. I will build on previous talks with banks about how to get the CEO’s to take action. The bottom-up approach does not work, so we need to focus on regulators to force banks to act. Will present his view on who, when, how long things could take, and why and what needs to be done in order to prepare for the Y2Q, and did people think about the unix end-of-time which happens in 2038? A lot of resources are needed in the coming years.

14:30 - 15:00

Quantum Algorithms for the Pathwise Lasso

We present a novel quantum high-dimensional linear regression algorithm with an $\ell_1$-penalty based on the classical LARS (Least Angle Regression) pathwise algorithm. Similarly to available classical numerical algorithms for Lasso, our quantum algorithm provides the full regularisation path as the penalty term varies, but quadratically faster per iteration on the number of features/predictors $d$ and the number of observations $n$ under specific conditions.

HEAR WHAT PEOPLE ARE SAYING